Calibrating Input Level for Plugins

So you end up with more redundant models, more scrolling, more time experimenting just to find a “sweet spot”, more doubts, less fun.

Fwiw, not directly related to this topic, but in case of that issue (which is a general issue having tons of modelers, regardless whether their hard- or software, component modeling or capturing, whatever...) I seem to be making great progress. I got sick of all the tone hunting and by now I just look for a handful of specific baseline sounds and go from there. In DAW land they're instantly saved as channel strip presets, so I don't even need to open the used plugins anymore. In hardware/live land I resort to 2-4 different baseline models and use some dirt boxes to accomodate the situation. Seems to be working just great.

I actually only got back into amp plugins because of that new Macbook Air (allowing me to run everything at very low RT latency), but I try to not get crazy. Seems to work just fine. The only things I added outside of the HX ecosystem so far are 3-4 NAM captures and I found out that I actually like some stomp boxes of Logic's Pedalboard plugin (which is astonishing as the amps are pretty bad, maybe with the exception of the Large Tweed). As far as anything else goes, it's standard DAW FX.

Anyhow, in general, I obviously agree with the premise(s) of this thread.
 
Fwiw, not directly related to this topic, but in case of that issue (which is a general issue having tons of modelers, regardless whether their hard- or software, component modeling or capturing, whatever...) I seem to be making great progress. I got sick of all the tone hunting and by now I just look for a handful of specific baseline sounds and go from there. In DAW land they're instantly saved as channel strip presets, so I don't even need to open the used plugins anymore. In hardware/live land I resort to 2-4 different baseline models and use some dirt boxes to accomodate the situation. Seems to be working just great.

I actually only got back into amp plugins because of that new Macbook Air (allowing me to run everything at very low RT latency), but I try to not get crazy. Seems to work just fine. The only things I added outside of the HX ecosystem so far are 3-4 NAM captures and I found out that I actually like some stomp boxes of Logic's Pedalboard plugin (which is astonishing as the amps are pretty bad, maybe with the exception of the Large Tweed). As far as anything else goes, it's standard DAW FX.

Anyhow, in general, I obviously agree with the premise(s) of this thread.
Just out of curiosity, because I've been contemplating getting an Apple Silicon notebook, how low can you go as far as latency's concerned on the Air? And what audio interface are you using (Helix?)?
 
Just out of curiosity, because I've been contemplating getting an Apple Silicon notebook, how low can you go as far as latency's concerned on the Air?

Pretty much as generally, hence: as low as the interface drivers allow you to. Obviously CPU usage goes up with lower buffer sizes, so you need a decent machine to run amp modelers at lowest buffer settings, but these days, most amp modelers aren't an issue in case you don't run fullblown kitchen sink presets on each track (which isn't exactly a good idea anyway, outside of the live business).

Anyhow, I'm using a Motu M2, bought this because I was reading good things about its latency performance and they turned out to be true. I can run Logic at 32 samples buffersize (running projects with 44.1 or 48 kHz) all throughout and the roundtrip latency is at 3.5ms, which is as good as some hardware units (not even considering FX loops, which add another roundtrip of latency). You can only get better values in case you're spending quite some more bucks, such as for an RME Babyface Pro. I'm still considering that as a potential "one day" purchase, but I'm actually quite happy with the Motu (some minor drawbacks, but I can live with them just fine).

And what audio interface are you using (Helix?)?

As far as the Helix or Stomp as interfaces go, forget about them. Their drivers are solid but latency is the worst in their class. And no, unfortunately that's no hyperbole but proven. At 32 samples (44.1kHz) latency is around 16ms. Which is horribly bad.
 
Last edited:
  • Like
Reactions: Elf
Just in case anyone owns the Slash Amplitube expansion and would like to have a ballpark on how to set their input gain for a more realistic response:



The main thing you have to do is ignore any clipping lights or advice you may have read about gain staging. -6dBu is your headroom target, best to boost digitally rather than obliterating your converters (unless you're using particularly weak pickups).
 
I’ll do more video examples, but it seems that IK’s JMP-1 model is set up for around 20dBu (i.e. 26dB lower than what the AFD needs).

Mark V and AD30 seem to be -6dBu, SJ50 0dBu, VH4 4dBu.

GUESSING IS SHIT but at least starting here they sound somewhat like you’d expect them.
Nice, so their official models seem to be all over the place??? Maybe that's why they don't feel like providing proper input values. 🤷‍♂️
 
Nice, so their official models seem to be all over the place??? Maybe that's why they don't feel like providing proper input values. 🤷‍♂️
Yep, and that’s basically why it matters more for AT5 than other software where a ballpark value will be close enough for everything. Not only are IK’s all over the place but it’s by huge amounts (where the input slider and even 20dB boost pedal doesn’t have enough range).
 
I guess they have taken the view that most people crank the input gain on the audio interface and that represents a close interpretation of the amps in the software?

Never really worked for me and I have likely not gotten the most out of some of the amp models as a result.

My preference is the same as you - to try and gain stage so the amp sim responds as it would in the real world with the interface gain at 0. Changed the way I use amp sims for the better in my opinion.
 
I guess they have taken the view that most people crank the input gain on the audio interface and that represents a close interpretation of the amps in the software?
If I cranked my input gain to -0.1dBFS, I'd still be too low for many amps, while for others it would be way too hot (check the JMP-1 for an AT5 model that prefers a quieter signal). I think generally, most of (especially the older) the IK amps are expecting -6dBu or 0dBu, but its really hard to decipher for certain amps. For 0dBu or -6dBu, its basically impossible to record a signal hot enough that isn't clipping.

I know that this topic can be made to seen as a bigger deal than it is, but it also blows my mind when doing this how when you get the level right, they suddenly sound WAY more in line with Helix/STL/Fractal etc. IMO they are generally not quite as good but it levels the playing field a hell of a lot.
 
From the recommendations here, I purchased Helix Native and I really like it.
As a Designer and CG Artist, I do like realistic looking UIs .. but .. Natives flat UI doesn't bother me at all. Imo, it has it's own pros.
What matters to me the most is the Quality and the Flow, that I can just dial in my input, read on what the amps/effects are about, fiddle a little, have my IRs loaded, and just play.

Along with Native, I do have high hopes with Genome as well.
I am curious if we have any inputs from Two Notes about what level Genome expects?
 
From the recommendations here, I purchased Helix Native and I really like it.
As a Designer and CG Artist, I do like realistic looking UIs .. but .. Natives flat UI doesn't bother me at all. Imo, it has it's own pros.
What matters to me the most is the Quality and the Flow, that I can just dial in my input, read on what the amps/effects are about, fiddle a little, have my IRs loaded, and just play.

Along with Native, I do have high hopes with Genome as well.
I am curious if we have any inputs from Two Notes about what level Genome expects?
Hi @abhijit_singh,

Amy from Two notes here! GENOME has been specified for levels in the range of -18 to -6dB, however we recommend normalising your input as best as possible for optimum results. Let us know if you have any questions concerning this!
 
Amy from Two notes here! GENOME has been specified for levels in the range of -18 to -6dB, however we recommend normalising your input as best as possible for optimum results. Let us know if you have any questions concerning this!
Appreciate you helping out with the discussion here Amy!

When people are asking about input levels, they’re generally looking for something more specific than just peak levels from the guitars. If you’ve seen the rest of the thread, there is information from pretty much every manufacturer making amp sim plugins giving their internal reference level relative to a 0.775V sine wave - i.e. if there is a constant analog sine wave that is 0.775V RMS, what dBFS level should this appear as in your software.

Each company uses different definitions for this spec, and I think if we’re only told “make your guitar peaks fall into this range” people will still continue to ask for a more specific value.

This thread is a bit of a goldmine on the subject and it’s a good thing to be aware of and understand IMO.
 
I've just got a Zoom AMS-22, and again, measuring it's maximum input gain, the result is 14.9dBu, and not the 11dBu stated in the manual.

My Audient EVO4 was even more deviated.

How can We trust manufacturers' tech specs?
 
I've just got a Zoom AMS-22, and again, measuring it's maximum input gain, the result is 14.9dBu, and not the 11dBu stated in the manual.

My Audient EVO4 was even more deviated.

How can We trust manufacturers' tech specs?
Replying to myself:

What´s accurate in both Audient and Zoom tech specs is the output level.

Zoom states -2 dBu, and I measure -1.9 dBu.

Audient states +11 dBu, I measure +10.9 dBu.

All good here.
 
Deciding whether to post this in a ToneX thread or input level thread.... anyway.
Finally made some strides on getting an actual good ultra high gain tone in ToneX. I'm talking to IK support with some observations and more questions etc... If anyone wants to jump in and take a look at this model I'd be interested to see how you find it on your end.


It's an EVH 5150iii RED channel and to get it sounding about right my levels are hitting it at +8dbu (My input is 13dbu and im gaining it up by 5db). The amp alone with that level of input gain is pretty competent but its still missing something. If you put another instance of ToneX before this with a tubescreamer all at unity then those two combined I'm finally getting a NeuralDSP level of gain out of ToneX.

Interested if it sounds the same as any old ToneX capture or you also think its come out top of the pile when it comes to gain.

FWIW my reamp out levels are 13dbu and the big difference is that I'm now ignoring the ToneX trainer when it comes to levels. I'm looking at the loadbox IN levels and setting it about 0.5db under clipping (basically as hot as absolutely possible). Once it's there I do the training and just hit next next next in ToneX, ignoring the guides.

1722434425665.png
 
Last edited:
  • Like
Reactions: Elf
Back
Top